Download 12.002 Physics and Chemistry of the Earth and Terrestrial Planets

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Large igneous province wikipedia , lookup

Age of the Earth wikipedia , lookup

Physical oceanography wikipedia , lookup

Geophysics wikipedia , lookup

Transcript
MIT OpenCourseWare
http://ocw.mit.edu
12.002 Physics and Chemistry of the Earth and Terrestrial Planets
Fall 2008
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
25
Planets as Heat Engines
In the earliest days of the solar system, all the terrestrial planets and
moons were exceedingly hot – most were probably entirely molten. Today
they are much cooler, although only the internal temperature of the earth is
well known. Sparse data are available for the mantles of the moon, Venus
and Mars, largely from the chemistry of igneous rocks now at the surface.
Rocks within the interiors of planets also contain various amounts of
radioactive elements, mainly uranium, thorium and potassium. The core
probably contains rather little in the way of these heat producing elements,
while the mantle contains much more. On earth, the crust is highly enriched
in radiogenic elements compared to the mantle. Heat is also generated by
freezing of the core, through the heat needed to turn fluid outer core into
solid inner core. The moon and Mars probably have solid cores. The states
of the cores of Mercury and Venus are unknown, although there are a few
data that bear on the question (magnetic field strength).
26
http://eqseis.geosc.psu.edu/~cammon/HTML/Classes
/IntroQuakes/Notes/Images_specific/Geotherm.gif
Courtesy of Prof. Charles J. Ammon, Penn State University. Used with permission.
Mechanisms for Planetary Heat Loss
Conduction (diffusion of heat through material)
Convection (advection is heat carried by movement of material,
called convection when movement is thermally driven)
Radiation (loss of heat through generation of electromagnetic
radiation at planetary surface)
Within planetary interiors, heat transfer occurs by convection and by
convection. Which is more efficient, and which operates where?
First, define heat flow (q): the amount of heat (joules) per unit time
(seconds) through a unit area of surface (square meters). Heat flow in the
earth is measured in units of mW/m2, noting that a watt=joule/sec. Typical
values for the outermost layer of the earth are between about q=40 mW/m2
and q=100 mW/m2.
27
Conductive heat flow is always proportional to the thermal gradient in a
material, with a proportionality constant called the thermal conductivity, K,
measured in units of W/mK:
qconductive = K (∂T/∂z)
Typical values for the thermal gradient in crustal rocks are between
10°C/km and 50°C/km.
Convective heat flow is computed as the velocity of the material moving
through a unit surface area, v, times the temperature of the material times the
heat capacity of the material. Heat capacity per unit volume is equal to the
heat capacity per unit mass (Cp) times the material density, ρ. Thus:
qconvective = v T ρ Cp
In general, convective heat loss in large bodies is much more efficient than
conductive heat loss, especially over distances of more than 100 km. This is
small compared to the radius of most terrestrial bodies, so it is difficult for
the planets to cool much by conduction. But, in order for convection to
occur, the interior of the planet has to be soft and weak enough that the “heat
engine” of the planet can drive ductile flow within the interior. Planets are
not hot enough in the upper tens to hundreds of kilometers for ductile flow
to occur, which needs temperatures somewhere above 1300°C even in the
shallow mantle. This means that all of the terrestrial planets have a strong
outer shell that does not deform in a ductile fashion (it could have faults).
Depending on the temperature of the planetary interior, there may or may
not be a ductile region in the deeper part of the planet’s mantle. This strong
outer layer is called the lithosphere (lithos means rock). It is not the same as
the crust. It usually includes all of the planetary crust and some of the
uppermost mantle. In the earth, the tectonic plates consist of lithosphere.
Mantle convection is driven by the density inversion in the mantle. All
things being equal, hot material is less dense than cold material. Thus
deeper and hotter parts of the mantle are more buoyant and tend to rise,
while shallower and colder parts of the mantle are less buoyant and tend to
sink. Although the details are a bit more complicated, this is the basic idea
behind thermal convection.
28
The mantle convects slowly moving a few centimeters in a year. The core convects more quickly, moving
kilometers in a year. The atmosphere convects vigorously, powered by incoming heat from the Sun.
Figure by MIT OpenCourseWare.
Courtesy of Walter S. Kiefer, Lunar and Planetary Institute. Used with permission.
http://www.lpi.usra.edu/science/kiefer/Research/convect4FS.gif
Math for Heat Conduction
29
Heat conduction, in one dimension, is governed by a partial differential
equation that has derivatives in time and depth. We are going to ignore the
effect of radiogenic heat production for our class work, although it can be
very important in crustal rocks.
Consider a layer of material conducting heat, with a heat flow out the top of
q and a heat flow in the bottom of q+dq.
The total heat in the box, Q, is equal to the temperature in the box, T, times
the heat capacity per unit volume, ρCp, times the volume (A dz):
Q = T ρCp Adz
Because everything is a constant except T, the rate of change of heat in the
box is just:
(∂Q/∂t) = (∂T/∂t) ρCp Adz
The rate of change of heat in the box is also equal to the heat flowing into
the bottom of the box, (q+dq)A, minus the rate of heat flowing out of the
top of the box, qA. So:
(∂Q/∂t) = (q+dq)A - qA
Setting the right hand sides of these equations equal we get:
(∂T/∂t) ρCp Adz = (q+dq)A – qA
or:
(∂T/∂t) ρCp = [(q+dq) – q]/dz = ∂q/∂z
30
Substituting the relationship between conductive heat flow, thermal
conductivity and heat flow:
q = K (∂T/∂z)
gives:
(∂T/∂t) ρCp = ∂/∂z [K (∂T/∂z)] = K (∂2T/∂z2)
or:
(∂T/∂t) = κ (∂2T/∂z2)
where κ= K/ρCp is the thermal diffusivity of the material, in units of m2/s.
This is the general equation for the conduction of heat in a stationary
medium without heat production and with uniform thermal diffusivity. In
the earth’s crust and mantle, a ballpark value for thermal diffusivity is κ=
10-6 m2/s, at least enough to make some approximate calculations.
A quick way of estimating how far, ζ, a significant change in temperature
can propagate in a given amount of time, τ, can be estimated by writing:
(dT/τ) = κ (dT/ζ2)
or
ζ = (κτ)1/2
putting in a variety of values for time (τ) and κ = 10-6 m2/s gives
characteristic conduction distances of:
τ
1 m.y.
100 m.y.
10 Gy
ζ
5.6 km
56 km
560 km
Since the planets are about 4.5 Gy old, cooling via thermal conduction can
not have penetrated more than a few hundred kilometers downward from the
surface. This means that if the planetary interiors have cooled at greater
depth, they must have done so by convection. If convection were to be
inhibited in the terrestrial planets, they could not cool significantly and
31
would remain near the same temperature, at least on the timescale of the
solar system.
Steady-State Conductive Solution
One simple solution to the heat conduction equation, which is obvious from
inspection, and is the “steady-state” or “equilibrium” solution when ∂T/∂t =
0. This gives:
(∂T/∂t) = 0 = κ (∂2T/∂z2)
or:
T = To + bz
The steady-state solution is just a linear function of z and corresponds to a
heat flow q = Kb. For now, we can make a table of the time it takes for a
given thickness of rock (mantle) to approach a steady-state linear geotherm
if we don’t perturb it in any way except to let it cool conductively.
time to steady-state
1 m.y.
100 m.y.
10 Gy
layer thickness
3 km
30 km
300 km
The Error Function
We will use the conduction equation in a few more classes to relate
lithospheric thickness to time and temperature so we might as well right
down one very useful time-dependent solution, called the error function and
written erf(y).
2
erf (y) =
"
y
%e
# y$ 2
dy $
0
This function has the property that erf(0)=0, erf(infinity)=1, and it is a
solution to the heat conduction equation. We will leave it to the homework
to show that it!
is a solution and to calculate the rate of conductive cooling of
a planetary surface. Later on we will use this expression to calculate the
thermal behavior of the oceanic lithosphere on earth.
32
Error Function
1
0.8
0.6
0.4
0.2
0
0
1
2
3